MC
M.C. Castelyns
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
1 records found
1
Data-Driven Control of Floating Offshore Wind Turbines
Using Multi-Input Multi-Output Closed-Loop Subspace Predictive Control
Floating Offshore Wind Turbines (FOWTs) provide access to deep-water regions with strong and consistent wind resources, but they also introduce additional control challenges compared to fixed-bottom turbines. In the above-rated operating region, the coupling between aerodynamic thrust and platform motion gives rise to non-minimum-phase behaviour, which limits the achievable control performance and complicates the design of effective controllers. Furthermore, obtaining sufficiently accurate control-oriented models for FOWTs is difficult because the relevant dynamics vary with operating condition and are influenced by the floating support structure. This thesis investigates closed-loop Subspace Predictive Control (SPC) as a data-driven control framework for FOWTs in above-rated operation. Two controller formulations are considered: a blade-pitch only Single-Input Single-Output (SISO) formulation and a Multiple-Input Multiple-Output (MIMO) formulation using both blade-pitch and generator-torque. The MIMO formulation is extended with an additional low-frequency torque penalty. This is introduced because slow variations in generator torque appear directly in the generated electrical power, whereas the main control objective in above-rated operation is power regulation. The penalty therefore suppresses undesirable low-frequency use of the torque channel while still allowing faster transient torque action. The proposed framework is first verified on a linearized floating wind turbine model. It is then evaluated on a high-fidelity nonlinear QBlade model to assess the controller behaviour under more realistic conditions and varying operating points. The results show that the proposed torque-shaping term can be incorporated naturally into the predictive-control formulation and effectively suppresses undesirable low-frequency torque actuation.
...
Floating Offshore Wind Turbines (FOWTs) provide access to deep-water regions with strong and consistent wind resources, but they also introduce additional control challenges compared to fixed-bottom turbines. In the above-rated operating region, the coupling between aerodynamic thrust and platform motion gives rise to non-minimum-phase behaviour, which limits the achievable control performance and complicates the design of effective controllers. Furthermore, obtaining sufficiently accurate control-oriented models for FOWTs is difficult because the relevant dynamics vary with operating condition and are influenced by the floating support structure. This thesis investigates closed-loop Subspace Predictive Control (SPC) as a data-driven control framework for FOWTs in above-rated operation. Two controller formulations are considered: a blade-pitch only Single-Input Single-Output (SISO) formulation and a Multiple-Input Multiple-Output (MIMO) formulation using both blade-pitch and generator-torque. The MIMO formulation is extended with an additional low-frequency torque penalty. This is introduced because slow variations in generator torque appear directly in the generated electrical power, whereas the main control objective in above-rated operation is power regulation. The penalty therefore suppresses undesirable low-frequency use of the torque channel while still allowing faster transient torque action. The proposed framework is first verified on a linearized floating wind turbine model. It is then evaluated on a high-fidelity nonlinear QBlade model to assess the controller behaviour under more realistic conditions and varying operating points. The results show that the proposed torque-shaping term can be incorporated naturally into the predictive-control formulation and effectively suppresses undesirable low-frequency torque actuation.